From 0db133ffb2b9841ba737329ce1a169954fe283b1 Mon Sep 17 00:00:00 2001
From: Paul Chaignon <paul.chaignon@gmail.com>
Date: Sat, 14 Mar 2015 19:43:17 +0100
Subject: [PATCH] Detect Cython generated files

---
 lib/linguist/generated.rb | 13 +++++++++++++
 test/test_blob.rb         |  4 ++++
 2 files changed, 17 insertions(+)

diff --git a/lib/linguist/generated.rb b/lib/linguist/generated.rb
index b93f7ea6..4a67fabd 100644
--- a/lib/linguist/generated.rb
+++ b/lib/linguist/generated.rb
@@ -62,6 +62,7 @@ module Linguist
       generated_parser? ||
       generated_net_docfile? ||
       generated_postscript? ||
+      compiled_cython_file? ||
       generated_protocol_buffer_go? ||
       generated_protocol_buffer? ||
       generated_jni_header? ||
@@ -270,5 +271,17 @@ module Linguist
       # VCR Cassettes have "recorded_with: VCR" in the second last line.
       return lines[-2].include?("recorded_with: VCR")
     end
+
+    # Internal: Is this a compiled C/C++ file from Cython?
+    #
+    # Cython-compiled C/C++ files typically contain:
+    # /* Generated by Cython x.x.x on ... */
+    # on the first line.
+    #
+    # Return true or false
+    def compiled_cython_file?
+      return false unless ['.c', '.cpp'].include? extname
+      return lines[0].include?("Generated by Cython")
+    end
   end
 end
diff --git a/test/test_blob.rb b/test/test_blob.rb
index 7c8fddc1..cc4243ea 100644
--- a/test/test_blob.rb
+++ b/test/test_blob.rb
@@ -240,6 +240,10 @@ class TestBlob < Minitest::Test
     # Godep saved dependencies
     assert sample_blob("Godeps/Godeps.json").generated?
     assert sample_blob("Godeps/_workspace/src/github.com/kr/s3/sign.go").generated?
+
+    # Cython-generated C/C++
+    assert sample_blob("C/sgd_fast.c").generated?
+    assert sample_blob("C++/wrapper_inner.cpp").generated?
   end
 
   def test_vendored